This repository serves as a curated collection of recent research papers and resources focused on Transformer architectures applied to Computer Vision (CV) tasks. It aims to keep researchers and practitioners updated on the latest advancements in this rapidly evolving field, providing a centralized hub for relevant literature.

How It Works

The project functions as a comprehensive, albeit irregularly updated, bibliography. It categorizes and lists papers, projects, and code repositories related to Transformers in CV, covering a wide spectrum of applications from image recognition and segmentation to video analysis and generative models. The inclusion of links to papers and code allows for direct access to the research.

Highlighted Details

• Extensive coverage of Transformer applications across various CV domains.
• Regularly updated with recent publications, including surveys and specific task-oriented papers.
• Provides direct links to research papers (arXiv) and associated code repositories.
• Includes foundational papers like "Attention is All You Need" and key models like CLIP and DALL-E.

Maintenance & Community

The repository is maintained by DirtyHarryLYL, with an open invitation for contributions and comments. Updates are described as "irregular."

Licensing & Compatibility

The repository itself does not host code or data, but rather links to external resources. The licensing of the linked papers and code would be specific to each individual project.

Limitations & Caveats

The primary limitation is the "irregular" update frequency, which might lead to a lag in capturing the absolute latest research. The repository is a curated list and does not provide implementations or tools itself.